Medical imaging plays a not insignificant role in personalized treatment decisions. Until now, medical imaging data was used primarily to derive geometric parameters (e.g., stroke volume of the heart, size of a tumor) to assist with a diagnosis. But this kind of imaging data contains a lot more information - called image-based biomarkers.
In oncology, for example, they can be used to more closely analyze tumors based solely on the imaging data. Instead of an operation to take samples of the tumor tissue, information can be obtained from tissue characteristics with the help of a virtual biopsy based on a quantitative image analysis. With this approach, called radiomics, the areas of radiological imaging data that contain a tumor are analyzed in depth, automatically extracting parameters such as the roundness of the tumor or the homogeneity of the gray value distribution inside of the tumor. This creates a high-dimensional feature space that can then be studied and visualized with visual analytics methods.
We continue to develop current radiomics approaches and apply them to entirely new fields. For example, we are currently working on a process for virtual biopsies of lymph nodes in ultrasound imaging data (echographs) that we, by analogy with radiomics, have christened ECHOMICS.